1

OX-5s to Turbo-Compounds: A Brief Overview of Aircraft Engine
Development

By Kimble D. McCutcheon

During the period between the World Wars, aircraft
engines improved dramatically and made possible
unprecedented progress in aircraft design. Engine
development in those days, and to a large extent
even today, is a very laborious, detailed process of
building an engine, running it to destruction,
analyzing what broke, designing a fix, and repeating
the process. No product ever comes to market
without some engineer(s) having spent many long,
lonely, anxious hours perfecting that product. This is
especially true of aircraft engines, which by their very
nature push all the limits of ingenuity, materials, and
manufacturing processes.

Aircraft Engine Requirements and
Measures of Performance

In order to compare engines, we must discuss the
special requirements of aircraft engines and
introduce some measures of performance. The
requirements are in some ways contradictory, and
therein lies the engineering challenge. For the
purpose of this discussion, we will compare the
Curtiss OX-5 to the Wright R-3350. The OX-5,
though hardly state-of-the-art at the end of WWI, was
the first U.S. aircraft engine to be mass-produced
and was produced in such quantities that war surplus
ones powered aircraft for the next twenty years. The
Wright R-3350, completely state-of-the-art at the end
of WWII, had been developed for the Boeing B-29
(the aircraft that dropped the atomic bomb on Japan)
and was widely used in airline service through the
middle sixties.

RELIABILITY

The first and most important requirement for an
aircraft engine is that it must be reliable. At the end of
WWI, the Curtis OX-5 regularly failed after only 30
hours of operation. During the 1950’s, airlines often
ran Wright R-3350s 3000 hours. This hundred-fold
increase in reliability is one of the fascinating
subjects of this discussion. These values are usually
expressed in Time Between Overhaul (TBO), but are
not really directly comparable. Pilots often ran OX-5s
to failure and forced landings were common. Airlines,
on the other hand, figured a forced landing might
scare their passengers, so they put on multiple
engines, kept good records about how long particular
engines could be expected to last, and presumably
overhauled them before they failed. The point,
however, is that engines got much, much better
during the period of our interest.

POWER-TO-WEIGHT RATIO

Secondly, aircraft engines must produce as much
power as possible while weighing as little as
possible. This is usually expressed in terms of
pounds per horsepower (lb/hp). One way to make an
engine more powerful is to make it bigger, but this
also makes it heavier. Moreover, if you shave away
metal to make it lighter, parts start to crack, break,
and generally become less reliable. You can see the
conflicting objectives faced by the engineer. Another
option is to get more power from a given size. Engine
size is usually expressed in cubic inches (cu in) of
swept volume (the volume displaced by all the
pistons going up and down). If you can make an
engine get more horsepower per cubic inch (hp/in),
then you have made it lighter. The OX-5 displaced
503 cu/in, weighed about 390 pounds and produced
90 HP (0.18 hp/in, 4.33 lb/hp). By contrast, the R-
3350 displaced 3350 cu in, weighed 3670 lb., and
produced as much as 3700 hp (1.10 hp/in, 0.99
lb/hp), improvements of six-fold in horsepower per
cubic inch and over four-fold in power-to-weight ratio.

FUEL CONSUMPTION

Finally, an aircraft engine must be fuel-efficient. A
great deal of the take-off weight of an airplane is
dedicated to fuel. So if one can make the engine(s)
more fuel efficient, less fuel must be carried to go the
same distance, and more bombs, passengers or
freight can be carried instead. Fuel usage is
expressed in terms called Brake Specific Fuel
Consumption (BSFC). This is the number of pounds
of fuel an engine uses per horsepower per operating
hour (lb/hp/hr). Fuel is measured in pounds because
a pound of fuel is always the same amount of fuel,
while a gallon of fuel at 100 degrees weighs less than
a gallon of fuel at 20 degrees. BSFC for the OX-5
was about .53 lb/hp/hr, while the R-3350 was about
.38 lb/hp/hr. If one could compare a ten hour flight
under similar conditions and power settings, one
would have to carry 371 pounds of fuel for the OX-5
verses 257 pounds of fuel for the R-3350, or a
savings of 114 pounds. This may not seem like much
of a difference, but again, it is an unrealistic
comparison because of the huge difference in the
output of the two engines. In reality, tens of
thousands of pounds of fuel were carried in the huge
transports of the 1950’s, and improvements in fuel
consumption made significant differences in overall
aircraft capability. Indeed, ocean-crossing airliners
such as the Lockheed Super Constellation and

2

Douglas DC-7 would have not been economically
feasible without the superb fuel consumption of
advanced engines.

Areas of Improvement

So how were these remarkable improvements made?
They were done by systematically improving seven
areas of engine design and construction:
Arrangement, materials, cooling, induction,
lubrication, fuels, and operation. Most of these are
necessarily interrelated, as we shall see.

In addition to engine improvements, there were also
important advances in aircraft and propeller design.
Perhaps the greatest engine-related airframe
advance was the development of the NACA cowl that
reduced the cooling drag of air-cooled radial engines
to levels that were competitive with liquid-cooled
engines. The greatest propeller advance was the
introduction in the 1930s of controllable pitch and
later automatically controlled constant speed.
Constant-speed propellers allow engines to produce
maximum take-off power by turning maximum RPM
due to fine blade pitch, and then cruise at efficient
lower RPM through the selection of a coarse blade
pitch.

We will now briefly discuss each of these areas of
improvement. Many of the engines have companion
articles that go into greater technical detail.

Figure 1. The NACA low-drag cowl

Figure 2. The variable pitch propeller

Arrangement

Engine arrangement refers to the organization of
multiple cylinders around the crankshaft. There are
really only two ways of doing this - to put them all in a
row along the length of the crankshaft, as in the in-
line engine, or to put them around a single throw of
the crankshaft like spokes in a wheel, as in radial
engines. For a long time, aircraft designers were
overly concerned with frontal area of engines,
because this had to be accounted for in the design of
the airframe, and produced drag. In-line, opposed,
and V-type engines provide the least frontal area
because cylinders are “stacked” one behind the
other. Unfortunately, any engine flexes as it runs and
must be stiff enough so that it does not crack its
components. This requires a very heavy crankcase
and crankshaft. The radial configuration avoids this
problem by having a short, stiff crankcase and
crankshaft.

3

Figure 3. Engine arrangements

Figure 4. Propeller reduction gearing

Over time, designers learned to stack multiple rows of
radial cylinders together, and since this had the best
power-to-weight ratio, it became the preferred
configuration for high-power engines. Advances in
cowl design all but eliminated any frontal area
advantage of the in-line and V-type engine. Many other
configurations were tried, but none ever equaled the
multi-row radial engine for power-to-weight ratio.

The Curtiss OX-5, Rolls-Royce Merlin (V-1610), and
Ranger V-770 are examples of V-type engines. There

are many examples of multi-row radial engines, with
the Wright R-3350 and Pratt & Whitney R-4360 being
the latest and most highly refined. There are also many
examples of opposed engines.

4

While engines are able to achieve higher power by
turning higher RPM, propeller RPM is limited by tip
speed. In order to remain efficient, propeller tips must
remain below the speed of sound. Otherwise, engine
power is wasted overcoming the excess drag of
propeller tips making shock waves and noise. The
logical answer to this paradox lies in the use reduction
gearing, allowing the engine to turn faster than the
propeller. Propeller reduction gearing was a feature of
the 1903 Wright “Flyer”, but it took a considerable
amount of work to sort out the details of reduction
gearing for high-powered radial engines, particularly
multi-row radials. Each power stroke of the engine
tends to slightly wind up the crankshaft. The propeller
resists this winding, or torsion. When the power stroke
subsides, the somewhat springy crankshaft unwinds
producing a phenomenon called torsional vibration.
This plagued early engines, was not very well
understood, and was generally fixed by resorting to
huge spur or helical-cut gears with massive teeth that
could resist the shock loads imposed on the reduction
gearing by torsional vibration. Later engines saw the
development of planetary reduction gears with very
close tolerances that mitigated some of the effects of
torsional vibration.

It all came to a head when controllable-pitch propellers
fitted to early Wright R-1820 Cyclones began breaking
propeller shafts. It turned out that the greater weight of
controllable-pitch propellers increased the effective
mass of the propeller and allowed vibrations of certain

frequencies to actually fatigue the propeller shaft until it
broke. The solution was to fit tuned dynamic torsional
vibration absorbers in the form of massive dynamic
counterweights loosely attached to the crankshaft so
they were free to move slightly in the plane of rotation.
Weight and pendulum length were calculated so that
the dynamic counterweight vibrated at the same
frequency as the power strokes of the engine, but out
of phase so as to cancel out the effect of the torsional
vibration.

Both the Wright R-3350 and Pratt & Whitney R-2800
encountered another vibration-related problem. These
were the first multi-row radials with nine cylinders per
row, and they too began breaking engine parts early in
development. The problem in this case was traced to a
different mechanism, but was still vibration related.
Radial engines with the master /articulating rod system
produce slightly different motions for each piston/rod
combination, and can never have perfect balance. This
becomes more of a problem as the number of
cylinders per row increases. The unbalance tends to
make the engine move in a circle in the same plane as
the cylinders. Because two-row radials have a two-
throw crankshaft, two such motions acting at twice
crankshaft speed tend to cause the engine to wobble
about its center main bearing. This wobble causes the
propeller change its plane of rotation, and eventually
fatigues the propeller shaft to the breaking point. The
solution is rotate correctly sized counterbalances at
twice crankshaft speed and in same direction as
crankshaft rotation.

Figure 5. Principle of tuned dynamic torsional vibration absorber

5

Figure 6. Second order counterbalance

Materials

An engine designer, always striving for low weight,
typically makes everything out of the lightest material
that is practical. This usually translates into the use of
aluminum for the bulky components (such as pistons,
cylinder heads, and crankcases) and steel for the
highly stressed components (such as crankshafts,
connecting rods, and gears). Over time, designers
created lighter and stronger alloys, developed ways to
harden materials so they lasted longer, and most
importantly, learned ways of forming metal
components so that the “grain” of the metal (metals
have grain just like wood) was correctly aligned to
handle the stresses imposed on the part. This process,
called forging, vastly improved the strength of almost
all engine components. Consider the strength of a
crankshaft carved from a single plank of wood. Though
the grain of the wood is in line with the bearing journals
of the crankshaft, the throws of the crankshaft would
be cut across the grain and would be quite weak. This
was the precise problem of early engines. Crankshafts
were machined from giant chunks of steel that had
been hot-rolled so that all the grain of the metal was in
one direction. The forging process takes a hot chunk of
metal and hammers it into roughly the final shape. The
metal grain is forced to conform to the final shape and
is much stronger. Nearly all engines made after 1920
used forged crankshafts, connecting rods, and pistons.
As forging processes became better understood and
huge hammer forges became available, larger engine
parts such as crankcases were forged. The Pratt &
Whitney R-1340 “Wasp” was the first American radial
to use a forged crankcase.

Figure 7. Etched connecting rod rough forging
showing metal flow lines

Further benefits were obtained by improving the art of
casting large chunks of aluminum. In the early days,
crankcases with integral cylinders could not be cast
because no one knew how to make such large
castings without flaws. In-line and V-type engines with
the cylinders separate from the crankcase could never
be as stiff as a single large casting, and consequently,
were heavier than necessary. The Curtiss OX-5 is an
example of a separate-cylinder engine while the Rolls-
Royce Merlin is an example of a one-piece block.

Figure 8. J-5 cylinder

Cylinder heads are another example of the progress of
the casting art. Compare the Wright J-5 “Whirlwind”
with the Pratt & Whitney R-2800. Each engine has cast
cylinder heads, but the fins on the J-5 are much further
apart and much less deep than those of the R-2800.
Considerable experimentation was required to perfect
these extremely complex castings, and much work was
required to produce the pattern and the mold for each

6

one. The result was an enormous increase in fin area
and better cooling. Later heads were forged, with their
fins cut by special automated machines. Not only were
the forged heads about twice as strong as the best
cast ones, but the fins could be deeper and closer
together, resulting in higher powers and better cooling.
Forged heads can be seen on the Wright R-3350.

Figure 9. R-2800 cylinder

As Pratt & Whitney began to extract more and more
power from their early engines, they began to have
occasional master rod bearing failures in the lead-
copper plain bearings originally used. A massive
amount of effort was thrown into experiments with
different bearing materials. Eventually, it was
discovered that a silver bearing plated with lead and
then indium had extremely good wear properties. In
the 1950’s, an airline returned one of these bearings to
Pratt & Whitney for rework after it had run over 7,000
hours. Pratt & Whitney returned it saying there was no
wear, approving it for continued service.

Finally, improvements in the materials and fabrication
techniques for valves made significant improvements
in the power and durability of engines. Most of this
work was done first at the Royal Aircraft Factory at
Farnborough, England and later at McCook Field in
Dayton, Ohio. Experimentation with simple single-
cylinder engines determined the best materials and
geometry for valves, guides and seats. The sodium-
cooled exhaust valve was also invented at McCook

field. This valve featured a hollow stem partially filled
with liquid sodium. As the valve opened and closed,
the sodium sloshed about, moving heat away from the
head to the stem of the valve. All Wright and Pratt &
Whitney radial engines use this style of exhaust valve.

Cooling

No debate was more heated in engine design circles
than the one over cooling. As with most heated
debates, neither side in retrospect knew what it was
talking about. The choices were liquid cooling, where,
as in automobile engines, the cylinders are surrounded
by a liquid coolant (usually water and anti-freeze)
which removes excess heat from fuel combustion and
is circulated to a radiator where it gives up this heat to
the air. Air-cooled engines, like lawn mowers, have
cooling fins on the cylinders, and give up their heat
directly to the air. The subject is complex, and it took
many years to sort it out completely (indeed, it may still
not be sorted out). In the early days, air-cooling was so
poorly understood that almost no one could make it
work at all, and certainly not for any high-power
applications. Liquid cooling at least allowed the
production of four or five hundred horsepower engines.
But these were unreliable engines. The Army, who in
those days had the luxury of flight over land, preferred
liquid cooled engines because of their lower frontal
area. The Navy, on the other hand, discovered that
fully twenty-five percent of engine failures were due to
failure of the cooling system, and declared that “Liquid-
cooled airplanes make about as much sense as air-
cooled submarines!”

During the twenties, air cooling became much better
understood, and high-power air-cooled engines
flourished to such an extent that all work on liquid
cooled engines ceased, and both the Navy and Army
had to pay premiums to attract any interest among
engine companies an liquid-cooled engines. The major
improvements were made at McCook Field, and
appear on all air-cooled engines since. Innovations
included an aluminum cylinder head with the valves set
at a very wide angle to allow plenty of airflow around
the exhaust port. A steel cylinder liner with machined
cooling fins was screwed and shrunk into this
aluminum head, resulting in a gas-tight seal between
the head and barrel. The exhaust valve was the
sodium-cooled variety discussed above. Nearly all air-
cooled engines have cylinders of this design (it first
appeared on the Wright J-5 “Whirlwind”).

7

Figure 10. Wright J-5 Cylinder Development. Note improved heat rejection at exhaust port.

Induction

Induction is the process by which fuel is mixed with air
and introduced into the cylinder. Engine power is a
function of the pressure at which induction occurs. By
forcing more of the fuel-air mixture into the engine at
higher pressure, impressive additional power can be
achieved. This process is called supercharging.
Superchargers are pumps that increase the pressure
of the fuel-air charge. In aircraft engines, these nearly
always take the form of centrifugal compressors.

Figure 11. Supercharger impeller and diffuser

Improvements in superchargers greatly assisted the
increased production of power, and also allowed the
engine to produce sea-level power at considerably
greater altitudes than non-supercharged engines. Early
superchargers were just “rotary induction systems”,

and served little purpose other than to assure equal
distribution of fuel to all cylinders. As engine
development progressed, superchargers became
better and better compressors by providing higher
pressure while consuming less power.

Figure 12. Single-stage supercharger

Supercharger design is a tricky business. Not only
must the supercharger be efficient to avoid wasting
engine power and excessively heating the intake
charge, but it must also have a pressure rise and
pumping volume that is carefully matched to the
engine it is a part of. The first American production
engine to use a supercharger was the Pratt & Whitney
R-1340 “Wasp”. All early engines used superchargers
from the same source - General Electric. By the
1930’s, it became clear to both Wright and Pratt &
Whitney that the GE superchargers were very
inefficient, and both companies established their own
in-house supercharger design teams. These designs
went on to set records for efficiency and pressure ratio.

As supercharger boost levels improved, the need
arose to tailor supercharger output to the engine power
and altitude. This was the reason for development of
two-speed and two-stage superchargers. The Pratt &
Whitney R-2800 in the F4U Corsair is an example of

8

the two-stage supercharger. The huge casting behind
the last row of cylinders is almost entirely a two-stage
supercharger. Output air from the first stage is ducted
to the second stage for further compression. An
intercooler, which is a sort of air radiator to cool the
compressed intake charge was often fitted to these
highly-boosted engines.

Figure 13. Two-stage supercharger

The huge induction system on big engines with high
boost pressures full of explosive fuel/air mix can be
blown apart by backfires resulting from improper
operator technique. This is one of the difficulties with
having the carburetor at the entrance to the induction
system. A more acceptable solution is fuel injection,
preferably directly into the cylinder. In this situation, the
induction system is just pumping air, so designers do
not have to worry about backfires, uneven mixture
distribution, and carburetor ice.

Figure 14. Direct fuel injection

Another type of supercharging that is very effective is
turbo supercharging. In this application, engine
exhaust velocity is used to drive a turbine which is
connected to a centrifugal compressor which rams
more air into the engine. The combined package is
called a turbocharger. A valve called the waste gate
controls turbine speed. The turbocharger has an

advantage of not robbing as much horsepower from
the engine as gear-driven superchargers do.

Figure 15. Turbo-supercharger with intercooler

General Electric built all of the turbochargers used in
World War II. All high-altitude bombers (B-17, B-24, B-
29) and many fighters (P-38, P-47) used turbochargers
to maintain full engine power up to an altitude of
eighteen to twenty thousand feet.

Figure 16. General Electric turbosupercharger

Near the end of World War II, someone got the idea to
harness the wasted energy in engine exhaust by using
the exhaust to drive a turbine that was coupled to the
engine crankshaft. This process is called turbo-
compounding. Although numerous engines had
experimental test programs with turbo-compounding,
only the Wright R-3350 Turbo Cyclone ever saw wide
service. Referring to Figure 18, notice the three large
pressure recovery turbines spaced equally around the
aft side of the engine. Each of these was fed by the
exhaust from six cylinders and contributed nearly 200
additional horsepower (600 total) to the engine output.
Another advantage of turbo-compounding is the
exceptionally good fuel consumption.

9

Figure 17. Turbo-compounding schematic

Figure 18. Wright Turbo-Compound 18 showing two of
three power recovery turbines

Lubrication

Early engines were lubricated with vegetable oils
usually castor oil. Castor oil was chosen because it
had nearly constant viscosity (resistance to flow)
across its temperature range, and because it coated
the metal surface well so that the lubricating film was
not easily scraped or washed away. It had the

unfortunate characteristic of turning to a gel after being
heated and then cooled. For this reason, it was and
still is used only in “total loss” lubrication systems such
as rotary engines, model airplane engines, and
outboards.

The introduction of high-quality mineral oils allowed re-
circulation of the oil (drastically reducing oil
consumption) as well as the production of greater
power by assuring that metal parts were separated by
a thin film of oil and never came in contact with one
another. To do this, the oil has to be able to resist
mechanical pressure, heat, the tendency to oxidize,
and the tendency to lose viscosity. Originally, only
straight mineral oils were used. In the 1950’s, additive
packages were introduced to make the oil “Ashless-
Dispersant” (AD). AD oils leave no residue when they
burn away (hence the ashless) and are formulated to
keep contaminants in suspension until the oil is
changed. Nearly all oil in use today is the AD type.
Eventually, synthetic oils with superior lubrication,
viscosity, and stability will probably replace mineral
oils.

Fuels

Of at least equal importance to all other areas of actual
engine development is the development of fuels.
During WWI, pilots noticed that gasoline refined from
Romanian crude oil, ran better than that refined from
California crude. After the war, an investigation of this
phenomenon revealed that “bad” gasoline caused the
engine to detonate. Detonation is a condition in which
the fuel-air mixture in the cylinder burns explosively
rather than smoothly. It was further discovered that
pure iso-octane, a gasoline constituent of a certain
molecular structure, was about the best that could be
had. Hence, the Octane rating system was born. Early
gasoline was between 25 and 50 octane.
Combinations of poor cooling, high compression ratios
(the ratio of cylinder volume at the top and bottom of
the piston stroke), and/or excessive supercharging
lead to detonation, often with disastrous results. In the
late twenties it was learned that the addition of tetra-
ethyl lead to gasoline drastically improved its octane
rating, so much in fact that it was better than iso-
octane. Fuels that test better than iso-octane are rated
with Performance Numbers (PN) These improved
fuels, often as high as 145 PN, allowed higher
compression ratios and higher supercharger pressures
which resulted in doubling or trebling of engine power.
It is interesting to note that the Allison and Rolls-Royce
engines used in WWII Allied fighters got about the
same horsepower from around 1700 cubic inches that
German engines got from 2600 cubic inches. This was
almost entirely due to use of 115/145 PN aviation
gasoline in Allied aviation engines verses the 80-90
octane German fuels.

10

Toward the middle of World War II, another technology
came on the scene that further improved engine take-
off horsepower ratings. This was Anti-Detonation
Injection, or ADI. ADI was simply a pump that during
extreme power conditions such as take-off, injected a
mixture of water and alcohol into the induction system.
The alcohol was primarily to prevent freezing of the
water. ADI greatly improved detonation margin, but
since it consumed large quantities of water (which is
heavy), it was typically only used during take-off or for
short times in combat.

Operation

The final area of improvement is that of actual
operation of the engine. When the R-3350 entered
service in World War II, it often did not run more than
100 hours before having to be overhauled. In airline
service, it would sometimes last over 3,000 hours. It is
true that the early R-3350s had design problems that
were fixed as the engine matured, but another
important factor was how the engine was operated.
The early engines were run very hard and very hot,
often overheated, flown by inexperienced crews, and
maintained by poorly trained mechanics. In airline
service, engines were treated very well, kept cool,
flown and maintained by experienced and competent
crew. They were also better instrumented and better
data was kept which allowed correlation between
operational practice and longevity.

One of the most useful instruments introduced during
the war was the torquemeter. This device measured
the amount of power actually being delivered to the
propeller and allowed the crew to select power settings
accurately and to lean the engine correctly to prevent
overheating.

Conclusion

By 1950, aircraft piston engines had reached their
pinnacle of development. They had become light,
powerful, reliable, and fuel-efficient. But they had also
reached their pinnacle of complexity and probably
power. It is doubtful that anything larger than the R-
4360 could have ever been cost-effective simply
because of the number of precision parts and amount
of maintenance required. Even the R-4360 was never
popular in commercial service because it typically
required many hours of maintenance for each flight
hour, and sophisticated fault diagnosis equipment to
boot. Cylinders larger than around 200 cubic inches or
producing more than about 200 horsepower were not
practical, and engines with more than about 28
cylinders were not practical. It follows that engines
larger than six or seven thousand horsepower were
also not practical. Around 1945, engineering effort at
the major engine plants began to turn away from piston

engines to engines with much greater potential for
development - jets.

12-16-99